Researchers Question LeJEPA's Scalability to Larger Datasets
Researchers including Google DeepMind's Luca Eyre question LeJEPA's scalability, noting its public weights were trained only on ImageNet-1k and there is no evidence the method scales to larger datasets.
2026-11-07 ~ 2026-11-07 · 2 related posts
- Researchers question whether LeJEPA's method scales beyond ImageNet-sized data — gabriberton · 2026-11-07
- giffmana questions LeJEPA: no scaling law evidence vs DINO — giffmana · 2026-11-07